I'm assuming you're using Lighthouse Studio. If so, please click Analysis + Analysis Manager. Then, choose HB as the Analysis Type. There is a "gear" icon to the right of that, that lets you select settings for your HB run. Click that Gear icon. On the left of the dialog that appears, you can select the Respondent Filter dialog, which lets you isolate which respondents to include in the HB run.
However, you should note that it isn't necessary to run HB separately by different groups to achieve a straightforward (Frequentist) test of differences between groups of people. You can just run HB generically including all respondents, which will give you utility scores and importance scores for each respondent in the output. Then, you can use those utility scores (preferably the normalized zero-centered diffs) and also the importance scores (which are already normalized, of course, to sum to 100) within a cross-tabulation tool like SPSS or your favorite cross-tab tool to conduct F-tests and T-tests for differences between groups of respondents.
For more details, including more sophisticated Bayesian tests, please visit: https://www.sawtoothsoftware.com/169-support/technical-papers/cbc-related-papers/2033-statistical-testing